This model began as a very basic study examining the relationship between team recruit rankings and team win totals. The graph below shows each team’s 247 Sports composite recruit ranking compared to the average number of wins that recruiting class earned over their 4 years at their respective schools. This was done for every Power 5 team between 2010 and 2018.
This study, while not originally intended for predictive purposes, actually proved to be a good foundation for building a predictive model. As shown in the graph above, team win totals and team recruit rankings had a R2 value of 0.288. From here, several other variables, outlined below, have been added to the model to improve its predictive ability.
The latest model has proven to be very successful in predicting team win totals. The model was used to predict team win totals during the 2021-2023 seasons – doing so within an average of 1.957 wins. This 1.957 value is less than the average difference between actual win totals and bookmakers lines across these 3 seasons (2.017). Had the model been used to place bets on O?U team win totals, 65.99% of bets would have won. This model was built using data from the 2016-2019 seasons and includes data for all current Power 4 teams, plus Notre Dame. A full breakdown for each season is included below.
| 2023 | |||
| Team | Predicted Total Wins | Total Actual Wins | Preseason Line |
| Alabama | 11.41 | 12 | 10.5 |
| Arizona | 7.45 | 10 | 5 |
| Arizona State | 5.34 | 3 | 5 |
| Arkansas | 6.14 | 4 | 7 |
| Auburn | 6.21 | 6 | 6.5 |
| Baylor | 6.00 | 3 | 7.5 |
| Boston College | 5.98 | 7 | 5.5 |
| BYU | 5.81 | 5 | 5.5 |
| California | 5.24 | 6 | 5 |
| Cincinnati | 4.92 | 3 | 5.5 |
| Clemson | 10.53 | 9 | 9.5 |
| Colorado | 3.16 | 4 | 3 |
| Duke | 6.94 | 8 | 6.5 |
| Florida | 7.98 | 5 | 5.5 |
| Florida State | 9.44 | 13 | 10 |
| Georgia | 11.74 | 13 | 11.5 |
| Georgia Tech | 3.94 | 7 | 4.5 |
| Houston | 4.89 | 4 | 4.5 |
| Illinois | 6.93 | 5 | 6.5 |
| Indiana | 5.10 | 3 | 3.5 |
| Iowa | 7.78 | 10 | 7.5 |
| Iowa State | 5.75 | 7 | 5.5 |
| Kansas | 6.42 | 9 | 6.5 |
| Kansas State | 7.75 | 9 | 8.5 |
| Kentucky | 7.30 | 7 | 6.5 |
| Louisville | 5.20 | 10 | 8 |
| LSU | 11.54 | 10 | 9.5 |
| Maryland | 6.49 | 8 | 7 |
| Miami | 7.72 | 7 | 7.5 |
| Michigan | 10.91 | 15 | 10.5 |
| Michigan State | 7.30 | 10 | 5.5 |
| Minnesota | 6.46 | 6 | 6.5 |
| Mississippi State | 5.44 | 5 | 6.5 |
| Missouri | 6.50 | 11 | 6.5 |
| NC State | 8.11 | 9 | 6.5 |
| Nebraska | 4.93 | 5 | 6 |
| North Carolina | 8.16 | 8 | 8.5 |
| Northwestern | 4.62 | 8 | 3.5 |
| Notre Dame | 10.77 | 10 | 9 |
| Ohio State | 11.61 | 11 | 10.5 |
| Oklahoma | 9.80 | 10 | 9.5 |
| Oklahoma State | 8.91 | 10 | 9.5 |
| Ole Miss | 7.14 | 11 | 6.5 |
| Oregon | 10.28 | 12 | 9.5 |
| Oregon State | 6.10 | 8 | 8 |
| Penn State | 9.98 | 10 | 9.5 |
| Pittsburgh | 7.41 | 3 | 6.5 |
| Purdue | 4.17 | 4 | 5.5 |
| Rutgers | 4.59 | 7 | 7 |
| SMU | 7.60 | 11 | 8 |
| South Carolina | 8.23 | 5 | 6.5 |
| Stanford | 5.23 | 3 | 3 |
| Syracuse | 4.00 | 6 | 6.5 |
| TCU | 9.32 | 5 | 7.5 |
| Tennessee | 9.14 | 9 | 9.5 |
| Texas | 9.61 | 12 | 9.5 |
| Texas A&M | 8.94 | 7 | 7.5 |
| Texas Tech | 6.44 | 7 | 7.5 |
| UCF | 5.77 | 6 | 6.5 |
| UCLA | 7.55 | 8 | 8.5 |
| USC | 9.49 | 8 | 9.5 |
| Utah | 8.41 | 8 | 8.5 |
| Vanderbilt | 6.03 | 2 | 3.5 |
| Virginia | 5.75 | 3 | 3.5 |
| Virginia Tech | 5.51 | 7 | 5 |
| Wake Forest | 6.44 | 4 | 6 |
| Washington | 10.15 | 14 | 9 |
| Washington State | 6.10 | 5 | 6.5 |
| West Virginia | 5.90 | 9 | 4.5 |
| Wisconsin | 7.48 | 7 | 9 |
| 2022 | |||
| Team | Predicted Total Wins | Total Actual Wins | Preseason Line |
| Alabama | 11.49 | 11 | 10.5 |
| Arizona | 5.40 | 5 | 3 |
| Arizona State | 7.29 | 3 | 6 |
| Arkansas | 7.50 | 7 | 7.5 |
| Auburn | 7.94 | 5 | 6.5 |
| Baylor | 8.25 | 6 | 7.5 |
| Boston College | 5.68 | 3 | 6.5 |
| BYU | 7.24 | 8 | 8.5 |
| California | 5.99 | 4 | 5.5 |
| Cincinnati | 9.60 | 9 | 9 |
| Clemson | 10.85 | 11 | 10.5 |
| Colorado | 5.11 | 1 | 3 |
| Duke | 3.90 | 9 | 3 |
| Florida | 7.40 | 6 | 5.5 |
| Florida State | 7.39 | 10 | 6.5 |
| Georgia | 11.40 | 15 | 10.5 |
| Georgia Tech | 4.47 | 5 | 3.5 |
| Houston | 8.90 | 8 | 9 |
| Illinois | 4.91 | 8 | 4.5 |
| Indiana | 5.59 | 4 | 4 |
| Iowa | 8.22 | 8 | 7 |
| Iowa State | 6.48 | 4 | 6.5 |
| Kansas | 4.05 | 6 | 2.5 |
| Kansas State | 7.12 | 10 | 6.5 |
| Kentucky | 7.54 | 7 | 8 |
| Louisville | 6.11 | 8 | 6.5 |
| LSU | 8.81 | 10 | 7 |
| Maryland | 6.72 | 8 | 6 |
| Miami | 6.65 | 5 | 8.5 |
| Michigan | 10.61 | 13 | 9.5 |
| Michigan State | 8.24 | 5 | 7.5 |
| Minnesota | 7.00 | 9 | 7.5 |
| Mississippi State | 7.69 | 9 | 6.5 |
| Missouri | 6.17 | 6 | 5.5 |
| NC State | 7.63 | 8 | 7.5 |
| Nebraska | 6.63 | 4 | 7.5 |
| North Carolina | 8.09 | 9 | 3 |
| Northwestern | 5.54 | 1 | 4 |
| Notre Dame | 8.97 | 9 | 8.5 |
| Ohio State | 10.98 | 11 | 10.5 |
| Oklahoma | 9.38 | 6 | 9.5 |
| Oklahoma State | 8.74 | 7 | 8.5 |
| Ole Miss | 8.72 | 8 | 7.5 |
| Oregon | 9.06 | 10 | 8.5 |
| Oregon State | 5.18 | 10 | 6.5 |
| Penn State | 9.79 | 11 | 8.5 |
| Pittsburgh | 7.36 | 9 | 8.5 |
| Purdue | 6.25 | 8 | 7.5 |
| Rutgers | 6.33 | 4 | 4 |
| SMU | 6.02 | 7 | 7 |
| South Carolina | 7.84 | 8 | 6 |
| Stanford | 6.15 | 3 | 4.5 |
| Syracuse | 5.04 | 7 | 5 |
| TCU | 5.52 | 13 | 6.5 |
| Tennessee | 8.80 | 11 | 7.5 |
| Texas | 7.94 | 8 | 8.5 |
| Texas A&M | 9.17 | 5 | 8.5 |
| Texas Tech | 4.28 | 8 | 5.5 |
| UCF | 9.01 | 9 | 6.5 |
| UCLA | 7.06 | 9 | 8.5 |
| USC | 6.46 | 11 | 9.5 |
| Utah | 8.04 | 10 | 9 |
| Vanderbilt | 3.75 | 5 | 2.5 |
| Virginia | 6.34 | 3 | 7 |
| Virginia Tech | 5.77 | 3 | 6.5 |
| Wake Forest | 6.51 | 8 | 6.5 |
| Washington | 7.20 | 11 | 7.5 |
| Washington State | 4.42 | 7 | 5.5 |
| West Virginia | 6.15 | 5 | 5.5 |
| Wisconsin | 8.88 | 7 | 8.5 |
| 2021 | |||
| Team | Predicted Total Wins | Total Actual Wins | Vegas Odds |
| Alabama | 11.57 | 13 | 11.5 |
| Arizona | 4.83 | 1 | 2.5 |
| Arizona State | 7.17 | 8 | 9 |
| Arkansas | 4.59 | 9 | 5.5 |
| Auburn | 6.84 | 6 | 7 |
| Baylor | 6.20 | 12 | 0 |
| Boston College | 6.02 | 6 | 7 |
| BYU | 7.32 | 10 | 6.5 |
| California | 6.04 | 5 | 5.5 |
| Cincinnati | 9.36 | 13 | 10 |
| Clemson | 11.39 | 10 | 11.5 |
| Colorado | 4.92 | 4 | 4.5 |
| Duke | 5.75 | 3 | 3.5 |
| Florida | 8.74 | 6 | 9 |
| Florida State | 6.89 | 5 | 5.5 |
| Georgia | 10.77 | 14 | 10.5 |
| Georgia Tech | 5.11 | 3 | 4.5 |
| Houston | 7.32 | 12 | 8 |
| Illinois | 3.45 | 5 | 3 |
| Indiana | 5.63 | 2 | 8 |
| Iowa | 8.06 | 10 | 8.5 |
| Iowa State | 6.38 | 7 | 9.5 |
| Kansas | 3.04 | 2 | 1 |
| Kansas State | 6.86 | 8 | 5.5 |
| Kentucky | 7.21 | 10 | 7 |
| Louisville | 5.84 | 6 | 6.5 |
| LSU | 10.96 | 6 | 8.5 |
| Maryland | 5.70 | 7 | 5.5 |
| Miami | 8.17 | 7 | 9.5 |
| Michigan | 9.00 | 12 | 8 |
| Michigan State | 7.53 | 7 | 4 |
| Minnesota | 7.17 | 9 | 7 |
| Mississippi State | 6.40 | 7 | 5.5 |
| Missouri | 6.62 | 6 | 7 |
| NC State | 7.70 | 9 | 6 |
| Nebraska | 6.40 | 3 | 6 |
| North Carolina | 8.92 | 6 | 10 |
| Northwestern | 6.11 | 3 | 6.5 |
| Notre Dame | 11.18 | 11 | 9 |
| Ohio State | 11.35 | 11 | 11 |
| Oklahoma | 11.74 | 11 | 11 |
| Oklahoma State | 8.39 | 12 | 0 |
| Ole Miss | 6.27 | 10 | 7.5 |
| Oregon | 10.05 | 10 | 9 |
| Oregon State | 4.61 | 7 | 4.5 |
| Penn State | 9.45 | 7 | 9 |
| Pittsburgh | 7.06 | 11 | 7 |
| Purdue | 5.30 | 9 | 5 |
| Rutgers | 4.10 | 5 | 5 |
| SMU | 7.06 | 8 | 6 |
| South Carolina | 4.66 | 7 | 4 |
| Stanford | 6.79 | 3 | 4 |
| Syracuse | 5.63 | 5 | 3 |
| TCU | 6.78 | 5 | 7 |
| Tennessee | 5.73 | 7 | 6 |
| Texas | 7.74 | 5 | 8 |
| Texas A&M | 9.34 | 8 | 9.5 |
| Texas Tech | 5.35 | 7 | 4.5 |
| UCF | 7.96 | 9 | 9.5 |
| UCLA | 7.31 | 8 | 7 |
| USC | 8.26 | 4 | 9 |
| Utah | 8.09 | 10 | 8.5 |
| Vanderbilt | 3.27 | 2 | 3 |
| Virginia | 6.54 | 6 | 6 |
| Virginia Tech | 7.14 | 6 | 7 |
| Wake Forest | 5.94 | 11 | 6.5 |
| Washington | 9.96 | 4 | 9 |
| Washington State | 6.60 | 7 | 6 |
| West Virginia | 6.33 | 6 | 6.5 |
| Wisconsin | 8.85 | 9 | 9.5 |
The key variables used to create the model are:
- Team recruit ranking
- Home field advantage
- Returning production
- Program status
- Head coach tenure
- Recent win trends
The primary variable used to create this model is team recruit rankings, per 247Sports. Team recruit rankings, weighted by class and including transfer players, were used to assign each team a “Recruiting Value”(RV) for each season. Using individual matchup data from 2016-2019, the delta between a team’s RV and their opponent’s RV was plotted against the outcome. This produced a trendline indicating a team’s chance of winning a game at a given RV advantage/disadvantage.
In addition to the RV delta, teams were given a home adder based on their historic winning percentage in home games. Additionally, Bill Connelly’s “Returning Production Rankings” were used to increase/decrease a team’s winning percentage based on their level of experience. Using the trendline above along with the home adder and returning production, teams were given a percentage chance to win every scheduled matchup. Summing these values produced a raw win total for each team (RW).
After reviewing the RW data, it was apparent that additional variables were needed to improve the model. Most notably, teams that had undergone recent head coaching changes were identified for improvement. Additionally, teams that had recent history of exceeding/falling short of expectations were commonly misrepresented by the model. Thus, a new variable was created.
Head coach tenure and recent season win trends were used to categorize teams into one of several different categories, each with a corresponding PS value. Teams with long tenured head coaches and consistent win totals year over year received a neutral PS value. Teams with head coaches in years 1-3 of their tenure seeing an increase in year-to-year win totals received a positive PS value (ex. 2024 Colorado). The opposite is true of teams with head coaches in years 1-3 of their tenure with decreasing year-over-year win totals (ex. 2024 Purdue). Teams with new head coaches received a negative PS value (ex. 2024 Boston College), with a more severe negative for teams that lost successful head coaches (ex. 2024 Washington).
Ultimately, this model combines the RW and the PS variable to create a comprehensive win prediction tool that has demonstrated successful results over the past 3 seasons. The data shown on the home page of this site are the results for the 2024 season. Preseason predictions as well as updated predictions based on 2024 results are shown.
Additionally, the model can be used to predict results on an individual matchup level. Over the course of the 2021-2023 seasons, the model accurately predicted the winner of 73.83% (1069-379) of matchups that included at least one Power 4 opponent.
Please note, it is not recommended that this model or any of the data on this site should be used for sports betting or gambling of any kind. Sports betting data is referenced as a means of demonstrating the accuracy of the model only.